#greshamslaw

dredmorbius@joindiaspora.com

An HN reader comments: I'd like a world where everyone consistently applies critical thinking to all sources of information.

That specifically Does. Not Scale.[1]
It fails two ways:

  1. Individuals suffer information overload, trust breakdown, and validation fatigue.

  2. Society finds itself with no common foundation of common shared facts and mechanisms. All points of view are asserted to be equally valid, expertise is entirely dismissed. Tribal beliefs are asserted as true (for Us) and invalid (if Them).

There is, I'll posit, a broad gulf between "verify everything" and "be prepared to question any belief". One varient of the latter is "strong opinions, weakly held". I'm not fully convinced this is a valid model, but it seems a good initial approach. It addresses both the need. to act in the moment, based on. partial information, and the realisation that this information and conclusions based on it may be faulty. The problem occurs when making decisions with no recourse --- betting the farm, burning the boats.

Otherwise, I lean strongly on Baconian, Pragmatic principles: our brains, both individually and socially, are sense-making organs, optimising for practical benefit. A challenge is that subject to perception, processing, and model-generation costs, complexity and rigour, though affording greater accuracy and precision, have enormous costs.

A manifestly false assumption of rational markets (and behaviour) theory is that information is free. It's not --- it has extraordinary costs, and model formation and coordination --- getting everyone on the same page --- are among the highest. We're constantly facing a complexity cost constraint (this is the essence of Gresham's law), in which a much simpler model is, under relaxed environmental selection, often more useful, as it permits discarding expensive perception, processing, and model transmission (education of the population). Which works fine until environmental selection mechanisms are increased.

What a trust, not in authority but in expertise, has to offer, suject to sufficient checks, is an efficient distribution of information, processing, and model formation. This is the ultimate aim of Baconian Science, expressed in the motto of the Royal Society: In nullis verba --- on the word of no person. Rather, it is justified trust in experiment, experience, integrity, and institutions, that gets you this.

Mind, the usual problem is that power-serving institutions have a staggering tendency to become self-serving and select not based on truth but on self-interest. Correcting for this tendency is the great problem of polity, commercial, social, justice, and moral systems.[2]


Notes:

  1. As I was composing this reply, another HNer in a different thread makes similar remarks: https://news.ycombinator.com/item?id=23450409

  2. This piece comes as I'm trying to wrestle another essay on truth and epistemology. Though it, so far, is winning.

#truth #belief #socialKnowledge #culturalKnowledge #RogerBacon #trust #belief #pragmatism #GreshamsLaw

dredmorbius@joindiaspora.com

Community or Features

I've been thinking of a Community or Features post over the past day or two after someone at G+ noted that most of the site-and-platform comparison discussions in the Migration forum were focusing over features. @John Poteet managed to provoke me into trying to thumbnail that here (which always seems to create a book....)

(Spoiler alert: This is in fact a fucking book.)


Starting a viable discussion forum requires a viable community. You can have the best tech, and if you've got a bunch of people who have nothing to say, or fight, or attack one another, or attack newcomers, etc., you're going to be stuck. The community is the base.

But with any community, as you grow, and especially if you grow through the general public, you will lose that initial base community's characteristics. You're going to regress toward the mean. And even if you don't, there comes a time when size alone triggers frictions and storms and rages.

If you can grow through a large selective and preexisting community, you can extend that day for a while. This is precisely what Facebook did, by starting at Harvard, extending through the Ivy leagues, then selective-admissions colleges, then general academic institutions. I'd said this for years, and finally discovered that danah boyd also recognises this, validating yet another of my views.

And Facebook wasn't the first to do this. The path had already been established by Usenet, starting about a quarter century earlier, also at selective-admissions universities. Both networks effectively leveraged the pre-screening functions of university admissions (and hiring) offices.

Doing that created a core base population of ... probably on the order of a million or so. However many undergrads there are in the US, times some penetration percentage.

Even if you do this, eventually you reach a point where adding more people starts creating problems. And solving problems has a name: technology.

Health

Cities have more or less the same problem: they grow up to a point, and then problems set in. Crime, poverty, vice, disease. Mostly disease.

Cities solved this problem, mostly starting in the mid-19th century, by creating public health and civil works departments and supplying fresh, non-shit-polluted water, and building sewers to take out the filth. A favourite chart of mine is the New York Department of Public Health and Mental Sanitation, showing net mortality from 1800 - 2007.

The year the health depart was founded was the year that mortality peaked, and it fell 85% of its total decline before 1930. Which is to say, before most of what we consider modern medicine existed. Yes, there was germ theory, antiseptics, and anesthesia. There was no penicillin, few vaccines, no organ transplants, few cancer treatments, and a whole host of other stuff we now consider "medicine". But on the whole, especially if you had access to what there was, and a decent diet, clean water, and fresh air, you lived ... very nearly as long as today. Alcohol, tobacco, automobiles, coal smoke, and lead in paint and petrol were the major derogatory factors.

Economists who study healthcare tend to come away with the view that almost all the spending is wasted. Robert J. Gordon's The Rise and Fall of American Growth, a look at the US economy from 1870 - 2017 largely concludes this, especially for the period 1970-2017.

We need sewers

We aren't Harvard. (Some of us may have gone there, but for the most part ... not.) There is some selectivity within this group, but it's fairly modest.

After the initial stage of community growth, the ability to not annoy the fuck out of one another, to keep rage-storms under control, and to basically sustain the health of the overall community, will be what matters.

And that is where the feature lists start to matter.

You break the chain of disease by stopping transmission routes and eliminating breeding grounds. Sewers and freshwater disrupt the fecal-oral route of cholera and typhoid. Eliminating mosquitoes drives down yellow fever and malaria. TB only eventually responded to antibiotics, though quarantine helps. And building a recognition to certain agents -- vaccination -- is effective against many virally-transmitted diseases. For environmental contaminants, the "infection" isn't biological, but still has a vector: reducing exposure to the insults to health listed above

(Aside: so much for the fucking thumbnail, sorry folks.)

Crusher

When Mastodon first started kicking off ... I started writing an earlier version of this essay. In the midst of all the "it's a kinder-gentler place" and "the community is so nice" discussion, I raised the point that we were tiny, there were few controls, and we'd not yet seen any injected discord.

Eris arrived in the guise of Wil Wheaton.

Yes, the Star Trek NG actor. And no, he wasn't the problem, at least not in the sense of him behaving outrageously and kicking up shit. For the most part, he just was. Posted some stuff. Paintings. Etc.

I'm not a huge fan, I have nothing against him. I know he was on Star Trek, I though his character was kind of annoying, the few times I'd .watched. But that was a character in a TV show. Which I, frankly, didn't care about either. About the only other exposure I'd had was Stand By Me, which I saw and liked when it came out, though it didn't really register that that was Wil Wheaton.

He's also written a few things, some recently, that impressed me. Those favourably incline me toward him.

But there's a group of people who felt differently, and they felt that very strongly. They'd felt it on Twitter, previously. Which had poor controls, largely negligent administration, and hence: a problem. And so Wil had used what few resources he had at his disposal to combat widespread violence, one element of which was a blocklist of Twitter accounts, mostly compiled by someone else.

(Who, it turns out, also had problems being harassed and brigaded.)

For various reasons, one group disproportionately represented on the list was the, or at least part of the, LGBTQ community. Which the Someone Else is (if I'm getting my story straight) a part of, and ran into some conflicts with, and ended blocking numerous people from that, and others those people associated with. Sort of a friends-of-an-enemy blocklist.

Wil used that blocklist.

He shared it.

Celebrities are not like you and me. They know people.

And more importantly, people know them. Lots of people. Lots and lots and lots and lots of people.

On the Internet, I have a reach of ... if follower counts are accurate and independent, maybe 3,000 or so individuals. And I suspect most of those are bots and sex-spammers. But that's a fairly large social crowd by ordinary standards.

Wil is known by millions. Many millions. Possibly hundreds of millions.

(And, should celebs start appearing in Diaspora space, Things will Change, Quickly and Drastically.)

And a fair number of them are on Twitter.

So, when Wil got hounded by a bunch of people and started deploying blocks just to get some breathing room, and then found a bigger and more effective blocklist that solved his problem, and then shared that list with others, some of whom were also likely well-interconnected on Twitter, and through whom tweets would, or wouldn't, get re-tweeted and seen ....

.... a bunch of folks who really hadn't considered their life choices, situations, vulnerabilities, and opportunities particularly well found that their ability to reach out and connect through Twitter, an ability some of them were relying on for at least part of their livelihood ....

Stopped dead cold.

And so they realised the error of their ways, that they'd done a pretty stupid thing, apologised to Wil, and emailed him at the address he'd set up for people to get off his blocklist, and for which he'd publicised and apologised for problems, etc., etc.

Yeah. Right.

No, being people, who don't consider their life choices, situations, opportunities, and vulnerabilities very well, some of them, at least, harboured a deep-seated and abiding rage that curled and roiled and grew and ... really couldn't do anything.

Until Twitter annoyed a bunch of people as it does from time to time, and a bunch of people looked up and saw this happy dancing elephant, and joined up with Mastodon, in late July and early August of this year. Including a bunch of claiming to be from the LGBTQ community, and Wil, who'd actually had an account for a while, but started using it again.

(And to be clear: not all of the LGBTQ community, and quite possibly numerous actors claiming a false flag to engender sympathy. Good friends who do identify with the community strongly criticised the event, once they came to see it for what it was.)

And they started needling him.

Repeatedly.

One Mastodon instance owner said, and I know this because I saw it and copied it to Archive.is and passed it on to Gargron, "you don't have to kill him, but" ... which is pretty close to saying "but you can if you want and I won't mind". And he did absolutely and unambiguously call for people to harass the shit out of Wil.

Mind, at the time, I didn't know most of what I'm describing above. Fog of war is a bitch, and what I saw was a whole bunch of folks claiming Moral Rectitude hounding a Hollywood Name of at Least Some Renown.

But it smelled kinda ugly and I dug a bit, and the story I'm telling above is more-or-less my sense of it.

And at the bottom of it was that most of them just seemed to hate Wesley Crusher. Who isn't even a real person.

But he was portrayed by one, and the hate transferred there.

60

So, Mastodon had a few fairly limited tools for dealing with problems.

Users can mute or block other users. They can block instances. And they can report problems to instance admins.

The entire Fediverse (Mastodon's protocol space) is somewhere around 1 million people. It's comparatively huge in Japan, which may be about half of that. Otherwise, the biggest instances with large English-speaking populations are mastodon.social, run by Gargron, Maston's creator, and mastodon.cloud, run by TheAdmin, whose real name name I don't know, though I know the handle, because that's my primary instance.

It's also the instance Wil was on.

So, people were reporting abuse. Some of them were reporting Wil, some were reporting the profiles attacking Wil. I reported the profiles attacking Wil.

Those reports go to the local administrator, and can optionally be forwarded to remote admins.

It turns out that, by sheer dint of size, a lot of the abusers were on Social and Cloud, though the highest concentrations (relative to server size) were on a small set of smaller servers -- a few hundred members each, possibly as much as thousand or so, I don't recall.

The admins of the larger servers were overwhelmed by the number of abuse reports coming in. And TheAdmin couldn't deal with the load, and so he direct-messaged Wil and told him Wil he was going to be booted off the network.

Now, I'm not sure if that was supposed to be a permanent boot or a temporary one, but ... Wil was the person being abused. He was the apple of discord, yes, but through no fault of his own.

Wil understandably got pissed. He'd been going through shit online for days at that point. Shit which from my own very limited experience warps your mind in weird ways (see my Imzy write-ups at Reddit and Ello). And even though he's got enough background as a celeb to know that people pull that kind of thing, it clearly got to him. He's written his side of the story, posted to his blog. And has, so far as I know, stayed off social media.

An overwhelming number of complaints. How many?

60.

Scale

As a tech-industry cat of a certain age and experience, I've worked at ... some levels of scale. Sites of 50 million users. Server farms with thousands to tens of thousands of machines (peanuts next to Google, mind, but at enough scale that You Have To Think Before You Do Stuff).

In a previous community migration I did, nearly 20 years ago, I'd realised that the crappy CMS forum software we'd been relying simply sequentially numbered posts, and I could walk through the archive from low to high and grab every item. A bit of binary searching found the high-numbered post. Then it was:

for i in [1..125000]; do 
    wget -o file-${i} "http://www.example.com/path/${i}"
done

A/K/A the good old bash one-liner. Count from 1 to 125,000, grab the url ending in that number, and output it to a file with the same number.

The resulting HTML was badly broken (part of the reason for the migration), in various ways, so I had to find out each of those ways and fix it. Almost all could be automated, so it was a case of track down each of a hundred or so cases of breakage and fix them.

I fetched and fixed the entire archive -- a few years' worth, in an evening or so, in my spare time, for the group who'd been using the forums.

Part of which may be a testimony to my obsessive nature, and part to the power of command-line tools. But also: 60 of something really should not be an insurmountable obstacle. And yet it was.

Maybe the UI was shitty. Maybe TheAdmin was having a bad day. Maybe he's not cut out for the social side of the gig (and truthfully, few are, it's brutal, especially when things are going poorly). Maybe he was up to his eyeballs in other issues.

But sixty of anything in a computer context, as an administrator, isn't worth blinking at, at least not in the sense of it stressing you out. Yes, the sixty things may mean there is a problem, but it's the problem that's the problem, not the sixty.

But sixty was the problem. As I understand this story.

Oh, and in case you think it's amateur hour that's the issue, on that 50 million user network, we managed to ... annoy our more vocal users one day (and yes, it was a moderately shitty thing we did, Because Business, though not World Ending). And even a Very Small Fraction of 50 million users can make an overwhelming noise if they're inclined to do so and you have no way of muting it. And that outfit sold for A Lot of Money at one point in time (and ... substantially less, not very long afterward).

What this means effective group hygiene means you've got to be able to operate effectively at scale.

If you don't, it's really simple: you won't scale.

Frogs

It's a bit worse than that though, in the online discussion space. Or, really, in any communications space. Because you don't just stand still. Things start getting worse, and often quite quickly.

There's a range of people and capabilities. Some are interested in and capable of intellectual and creative discussions, some aren't.

And it's not that one or the other is better or that if you belong to the Select Group of People with dots on their left earlobes or an outward-bending pinkie finger, You are Better than Them. It's just distributions.

And like good basketball players or musicians or hikers or programmers or artists or participants of any other activity frequently like to get together and riff off each other, you get a similar line with discussions.

And if there's a group that's keeping that from happening, it's frustrating. And eventually, the people who like to riff together go off somewhere quieter so that they can do that (if they can find one another, and agree on the same place, and have the time, and can Get to wherever There is with whatever Transport is available. (A windy way of alluding to different computing platforms, protocols, clients, servers and all that very non-musical jazz. Oh, and some goddamned thumbnail, right?)

So not only don't you scale but you start Driving the Good Bits off. Bad chat drives out good. Bad money drives out good. Bad politicians drive out good. We call it #GreshamsLaw, and anyone who's been around me for the past few years knows it's one of my obsessions.

Though the first observation came from the Greek playwrite Aristophanes in a play titled "The Frogs".

My Kingdom for a Thumbnail

Sorry to wear on, but I think I might be able to re-sketch this:

  • You need a good community to start.
  • That only gets you so far.
  • Then the Stink sets in.
  • If you had a good community, it Leaves.
  • Technology can ... if not avoid this, at least raise the size limit. Where "technology" is "tools for picking up and disposing of shit".

Wil left.

I've left forums. A bunch of folks left Google+, long before this past weeks' announcement.

They've also left Facebook and Twitter and who knows what else.

So it's not an either-or. But you need to get the balance right, and, if your group is looking for a platform and home, the either, it has to already have the tools you want, or it has to be capable of acquiring those. Which mostly means the dev team and others who can control direction of development (say, host admins who decide on comms protocols and codes of conduct) have to Not Get In The Way. Which happens a lot.

Mastodon's working on its problems. Things simmered down after the August blow-up. I don't think they're out of the woods. Federation of and by itself is not a silver bullet.

And so I'm looking for tools that offer promise of that forward path. It's very much on my mind as I'm contemplating the platform, site, service, and other choices in front of us.

How's that for a thumbnail?

dredmorbius@joindiaspora.com

Is Gresham's Law a special instance of a more general phenomenon?

I've just encountered three alternate presentations of Gresham's law in different contexts, and am beginning to convince myself that it's a specific case of a general type of systems failure.

Herman E. Daly describes a "Gresham's Law of Internalization":  with free trade and free capital mobility come a standards-lowering race.  Bad cost accounting drives out good cost accounting.  Ricardo's comparative advantage of trade assumed immobile capital....

Continued at the dreddit

#economics #greshamslaw #systems #systemsfailure #market #marketfailure

http://www.reddit.com/r/dredmorbius/comments/2h63fp/is_greshams_law_a_special_instance_of_a_more/